Representing System by Gene Encoding.

247
International Journal of Research and Reviews in Computer Science (IJRRCS)

Vol. 2, No. 1, March 2011

Representing System by Gene Encoding
Ahmad Yusairi Bani Hashim
Department of Robotics & Automation, Faculty of Manufacturing Engineering,
Universiti Teknikal Malaysia Melaka,
Hang Tuah Jaya, Durian Tunggal, 76100 Melaka, Malaysia

Abstract: The chromosomes that belong to the parents store
genetic information. The child inherits the information through
uniquely selective procedures. Using the concept, this study
proposes an approach to system representation using the process
found in genetic information inheritance. A generic process flow to
building the representation is shown in the form of ladder-like flow
chart that has the starting point at the bottom. This forms the basis
for gene encoding. Following this process flow, one can expect to
represent a system which requires solutions from selective
procedures. Therefore, the codes written may vary depending on

how one models and interprets the representation.
Keywords:
procedure.

Genetic

information,

chromosome,

selective

steps begin with creating population from data obtained
follows with the number of digits set for each chromosomes,
defines gender and set union rules.
Further, the number of offspring obtainable from a family
and the rules for transmission of genes from one generation
to another are set. There could be possibilities to adopt
offspring in case the family is not fit to produce one.
The rules for the most qualified offspring is established and

how they relate to their ancestors is investigated. Lastly, it is
to see if the selected ancestors become the sample that
possess the best chromosomes. The follwing section will
explain the steps depicted in Fig. 1.

1. Introduction
The genetic information that comes from the chromosomes
form strings that define a particular solution. If 110011
represent a chromosome, then a population of these
chromosomes corresponding to a number of individual
solutions to the problem is created.
There will be possibly some good solutions given by a
population of chromosomes. The next process is to measure
the fitness of each population of chromosomes. In addition,
the selected fitness function is specific to the particular
application, but generally it has a positive value when the
solution is good, whereas small value represents a bad
solution.
For example, if parent number 1 has chromosome 110011
and parent number 2 has chromosome 101010, then by

crossing over some of the genetic materials there exist
offspring number 1 chromosome 110010 and offspring
number 2 chromosomes 101011.
This crossing over is called the single-point crossover. The
process in crossing over actually is the breeding of a new
population. While a new population is being created,
mutation is allowed. A mutation is defined by the mutation
rate. For example a chromosome of a new population is
1110010101010 can mutate by a new chromosome
1110011101010. The bold digit changed from 0 to 1 showed
mutation has occurred.
This concept is generally accepted in genetic algorithm. In
this paper, however, we introduce a concept of genetic
information inheritance. It should be suitable for representing
a system which requires solutions from selective procedures.

2. Gene Encoding
The diagram in Figure 1 explains the steps in realizing the
gene encoding (GE). It has eight significant processes. The

Figure 1. The steps taken in realizing the GE. The sequence
is read bottom-up. From the initial sequence, the sequence
begins with raw data assembly and follows by the
determination of data population.

3. Representation
A population is created based on the type of input data.
Every population is unique. There will be no one population
that shares a type of input data. If there is more than one
input types then the population of the corresponding inputs
may be created.
In short, there will be a number of different sets of
population. For example, when in a system there are two
types of input data: angle, r and position, p . In this case,
there are two different populations of r and p respectively.
Input data can be described as a function such as f ( x ) ,
where x is the input data.
A population is then converted into strings or chromosomes.
The method for converting the populations to chromosomes
can be done in many ways. For example, the chromosomes

can be obtained by defined encoding schemes where a
permutation of all genes becomes a genetic chromosome [1].
A series of operation is used to represent a chromosome
such that in the operation that yields to scheduling of jobshop and flow-shop in a manufacturing system [2]. In
addition, chromosome can be represented as genetic string

248
International Journal of Research and Reviews in Computer Science (IJRRCS)

by applying objective functions of x and y where through
parameter coding [3]. One-half of the genetic string is
resulted from coding of x and another half comes from
coding of y. The length of a genetic string is to be defined
and optionally the size of a chromosome.
Researches in genetic algorithm typically related to
engineering application that define gender for particular
chromosomes in a population have not been found elsewhere
except in [4] where the work claims that all chromosomes
are of both sexes. Clownfish, for example, is naturally able
to change gender.

Obviously, a population that is created from input data is a
living entity. It is of our full control to assume the kind of
organic structure that is desired. Either the chromosome is of
male or female or both sexes depend largely on the end
results expected.
To have offspring there must be marriages. Unless there is a
certain orientation of chromosomes that has been defined to
represent a gender then the process for blending can take
place—the union. In [3], the cross-over process is applied to
produce offspring. There are two offspring resulted from the
crossovers. However, any information containing different
types of gene would generate illegal offspring [2].
The parents are chosen randomly and by partial cross-over
two offspring are produced. Could there be twins? One can
define that for each union a certain number of offspring are
allowed.
If a person possesses two of each chromosome, and two of
each genes type on each of a chromosome, he or she can
have two genes for eye color [5]. When both genes of a
given pair of chromosomes are identical, the person is said to

be homozygous for that gene.
On the other hand, when the two genes are different, the
person is heterozygous. There are genes called dominant
genes and recessive genes. Dominant genes are able to exert
their effects on development even in a person who is
heterozygous for that gene. Recessive genes possess by those
of homozygous’ only. A heterozygous’ will show the effects
of the dominant gene but still pass the recessive genes to a
son or a daughter.
It is necessary to have rules for transmission of genes from
one generation to another. In some applications there might
exist the requirement for producing the heterozygous
offspring.
There is likely to occur a situation where offspring are not
produced but the new population must expand. If such a case
occurs then offspring may be adopted. The orientation of
genetic strings of an adopted offspring must have a strong
resemblance to its original parents. The genetic conditions in
adoption process are, however, of the designer’s full control.
In a natural selection, in Charles Darwin’s theory of

evolution, is that individuals with certain genetically
controlled characteristic reproduce more successfully than
others. The artificial selection found in animal and plant
breeders is used for generating new strains. Moreover, a
person who has five healthy children before dying at age of
thirty is evolutionary successful [5]. For the new population
created, the qualified ancestors or the original parents can be
said to be evolutionary successful because their unions have
successfully produced offspring with unique genetic strings
patterns. Consequently, the qualified ancestors are said to

Vol. 2, No. 1, March 2011

have the most eligible chromosomes.

4. Model
We assume that every test would result in unique codes. A
test is called candidate, C . The sum of C becomes the
population of a culture. Equation (1) exhibits how GC is
represented. The symbol ∩ does not portray intersection or

union as in theory of sets. It describes the process of
matching C or the marriage, M . Only if a couple that has the
same genetic codes, then they can have a child, ch .
Equation (2) explains any union of C can only result a child.
It is agreed that any unions that produce a ch reflects to the
best combination of ancestors. The ancestors are those who
start the population. Any unions that do not produce a ch
are considered as the unfit ancestors, hence are abandoned.
This is summarized in Figure 2 where three candidates X ,
Y , and Z that are producing their successors.


1xxxx if exist 

GC =  gi ∈ bit gi = 
(1)

0xxxx otherwise 







1 qualified ancestor 

M k =  Ci ∩ C j } ∈ bit mk = 

0 unfit ancestor 




{

(2)

Figure 2. While Figure 1 is the general flow for GE this flow
chart, however, depicts of the processes executed using eqn.
(1) and eqn. (2).

5. Discussion
The GE shown in Fig. 1 is a general representation and is
open for modification. The common approach to visualize
this diagram is by a typical flow chart seen in Fig.3. We
avoid this type of chart because it is exact and the blocks
consume space. In addition, it stores information by the
shape of the blocks, which we think is specific and fixed.

249
International Journal of Research and Reviews in Computer Science (IJRRCS)

Vol. 2, No. 1, March 2011

though procedures in population determination and after the
union rules have been executed. In fact, a chromosome can
be added if there is a need for offspring adoption. The
inheritance is based on indefinite cross-over.

6. Conclusion
The method has its strength in analyzing sets of data and
gives solutions from selective procedures. This procedures
are predefined by the designer. It is of the designer’s total
control on how algorithms are arranged and manipulated.
Therefore, the steps in building the codes discussed in this
paper can be made as mere guidelines for specific software
design.

Acknowledgement
Thanks to Universiti Teknikal Malaysia Melaka for
providing sufficient resources and facilities for research
activities.

References

Figure 3. Another way of describing the GE as opposed to
Fig.1.
Our proposed GE is straight forward but is flexible. The
flexibility is found in the gender definition, union rules, gene
transmission rules, and offspring adoption. In gene
expression programming (GEP), however, the constrains of
the head-tail mechanism contribute to the legality of
chromosome [6]. In GE, legality of chromosome is based on
the procedures defined in the union rules and the adoption of
offspring.
On the other hand, cluster analysis is a major method to
study gene function and gene regulation information for
there is lack of prior knowledge for gene data. Many
clustering methods existed at present usually need manual
operation or pre-determined parameters, which are difficult
for gene data [7]. In GE, however, there is no clustering of
data except during the process of population determination.
This operation can be set manual or automatic depending on
the type of applications. If the application is distinct and
occasional then the operation is manual. Otherwise, it can be
set automatic if the application is routine and repeating.
Normally, in GEP, each chromosome is expressed and
evaluated in on the Expression Tree (ET). The ET-based
expression and evaluation are computationally expensive and
the intelligibility of the chromosome is low [8]. In GE, the
chromosome is not evaluated on the ET. It exists simply

[1] W. Wang and P. Brunn, P., “An effective genetic
algorithm for job shop scheduling,” Proc. Institution of
Mechanical Engineers, 214B, pp. 293-300, 2000.
[2] L.P. Khoo, S.G. Lee and Yin, X.F., “GA-based decision
support systems in production scheduling”, The Int. J. of
Adv. Manufacturing Technology, 8, 131-138, 2000.
[3] H.S. Ismail and Hon K.K.B. “The nesting of twodimensional shapes using genetic algorithms,” Proc.
Institution of Mechanical Engineers, 209, pp. 115-124,
1995.
[4] A.Y. Bani Hashim,
“Development of artificial
intelligent techniques for manipulator position control,”
Masters Thesis, Universiti Putra Malaysia, 2003.
[5] J.W. Kalat, “Introduction to psychology,” Third edition,
Brooks/Cole Publishing, California, 1993.
[6] J. Zhang, Z. Wu, Z. Wang, J. Guo, Z. Huang,
“Unconstrained gene expression programming,” Proc. of
the IEEE Congress on Evolutionary Computation, pp.
2043-2048, 2009.
[7] G. Liu, X. Jiang, L. Wen, “A clustering system for gene
expression data based upon genetic programming and
the HS-Model,” Proc. of the third International
Conference on Computational Science and Optimization
(CSO), pp. 238-241, 2010.
[8] Y. Chen, C.J. Tang, R. Li, M.F. Zhu, C. Li, J. Zuo,
“Reduced-GEP:
Improving
gene
expression
programming by gene reduction,” Proc. of the 2nd
International Conference on Intelligent Human-Machine
Systems and Cybernetics (IHMSC), pp. 176-179, 2010.